Rationale
Non-tumoral portal vein thrombosis (PVT) affects 5–26% of liver transplant (LT) candidates. It is known that PVT complicates LT surgery and impacts post-LT outcomes, and hence anticoagulation (AC) is recommended by all guidelines. However, limited data exists on the effectiveness of post-LT AC to prevent PVT recurrence and a uniform policy among transplant centers is lacking.
Objectives
1) Determine whether post-LT AC in patients with preexisting PVT prevents post-LT PVT and/or improves outcomes
2) Identify which patients will benefit most from AC.
Study design
International multicenter retrospective observational study in centers within the network of the Vascular Liver Disease Interest Group (VALDIG) and the European Liver and Intestine Transplant Association (ELITA).
Material and Methods
Inclusion criteria: adult (≥18 years) LT recipients who underwent first liver transplantation between 2010 and 2024 for liver cirrhosis and who had PVT at time of LT.
Data collection: Clinical and imaging data will be extracted from institutional records and recorded in a specifically designed CASTOR database. The (limited) dataset includes recipient demographics, LT indication, extent and treatment of PVT pre-LT, basic donor data, surgical details, use of post-LT anticoagulation, postoperative PVT, bleeding complications, graft and patient survival.
Primary outcome: post-LT PVT
Secondary outcome: anticoagulation safety (major bleeding), graft and patient survival
Statistical analysis
Results will be compared between 2 groups: patients who did and did not receive post-LT anticoagulation. Survival outcomes—including PVT-free, graft, and patient survival—will be evaluated using Kaplan–Meier methods and Cox proportional hazards regression, adjusting for relevant covariates. Missing data will be addressed through appropriate imputation methods, and propensity score matching will be applied if required to minimize confounding.
Sample size calculation
Based on a reported 3% overall incidence of post-LT PVT and a 5-fold increased risk of post LT PVT in those with pre-LT PVT, the estimated incidence of post LT without anticoagulation is assumed as 15%. 1) assuming a clinically relevant 50% reduction of this risk with anticoagulation (i.e. incidence 7.5%), with alpha set at 0.05, and a power of 80%, we will need 277 patients in each arm (i.e. 554 total). 2) more conservatively, a 3-fold increased risk of post-LT PVT would require 487 patients in each arm (i.e. 974 total). This emphasizes the need for large, multi-center data.